Pulling Weights from Formula

2 次查看（过去 30 天）
Conor O' Sullivan
Conor O' Sullivan 2014-12-15
评论： Sean de Wolski 2014-12-15
Hi all,
I created a relatively simple ANN using the pattern recognition wizard. 34 inputs, 2 outputs.
After creating the function, I wanted to pull the weights out to be used elsewhere (in an app), however I'm having some difficulty telling which numbers are the input weights, as well as the bias weight.
I've attached a word doc with the formula on it.
At first glance, I thought it might be the numbers next to x1_step1_xoffset or x1_step1_gain, however I'm doubtful they are them, as there's no negative numbers, and that would mean a positive result for any inputs whatsoever.
Anyway. I'd really really appreciate if anyone could shed some light on this for me.
Thank you,
Conor.

请先登录，再回答此问题。

回答（1 个）

Conor O' Sullivan
Conor O' Sullivan 2014-12-15
It's strange - when I try to attach, nothing happens. Here it is...
function [Y,Xf,Af] = myNeuralNetworkFunction(X,~,~) %MYNEURALNETWORKFUNCTION neural network simulation function. % % Generated by Neural Network Toolbox function genFunction, 12-Dec-2014 11:13:20. % % [Y] = myNeuralNetworkFunction(X,~,~) takes these arguments: % % X = 1xTS cell, 1 inputs over TS timsteps % Each X{1,ts} = 34xQ matrix, input #1 at timestep ts. % % and returns: % Y = 1xTS cell of 1 outputs over TS timesteps. % Each Y{1,ts} = 2xQ matrix, output #1 at timestep ts. % % where Q is number of samples (or series) and TS is the number of timesteps.
%#ok<*RPMT0>
% ===== NEURAL NETWORK CONSTANTS =====
% Input 1
x1_step1_xoffset = [17;34.3;84;40;45;8;84;0;0;0;0;0;35.1;87;37;41;9;87;0;0;0;0;0;35.1;81;35;47;9;87;0;0;0;0;0];
x1_step1_gain = [0.0298507462686567;0.307692307692308;0.0165289256198347;0.0285714285714286;0.027027027027027;0.105263157894737;0.125;2;2;2;2;2;0.408163265306123;0.0175438596491228;0.0263157894736842;0.028169014084507;0.133333333333333;0.153846153846154;2;2;2;2;2;0.425531914893617;0.0170940170940171;0.0240963855421687;0.0289855072463768;0.133333333333333;0.153846153846154;2;2;2;2;2];
x1_step1_ymin = -1;
% Layer 1
b1 = [1.6250159514304179;-0.76711617733625925;-0.087218477528884508;0.40218675522789599;1.5211715966196837];
IW1_1 = [-0.38334115088528592 0.073963977730160127 -0.21414592394085583 -0.1811023968674037 0.091457347558877411 0.26824792001744951 0.021515159445345867 0.16630488521234027 -0.31905356103014426 -0.028778051081310534 -0.2610524619567397 -0.3434506446145123 0.16255364446948642 0.46055087720622129 -0.52569637417996107 0.45356034166127779 0.41586228437989214 -0.042260852265273173 -0.09176126520559684 -0.1292117090529282 0.25370301418322144 -0.25521448463169621 0.22813842425566599 0.24812859798817519 0.33548716094810332 -0.28903101248327934 0.055823057258466077 0.20678256900006523 0.043714833424586721 0.25493727155065277 -0.2918105439897527 -0.085587139057542394 -0.026755880218446271 -0.19839458157068623;0.32344354450384089 -0.09722083870675273 -0.12141348409292725 -0.32147456751550335 0.22741347963361769 0.1866661505925552 0.11671432999224259 -0.1681892036000277 -0.079265244168909677 -0.32189673098914423 0.38782220143236523 -0.19557839247970041 0.10938064072960328 0.39493287272386546 -0.0070560903078537972 0.00061630728116105643 -0.20244959649453365 0.12395735927038111 -0.27396595888505826 0.35967072432173774 -0.40672023335160373 -0.039297405457635499 -0.24197601565354673 0.12517596889387095 0.0072890728100467317 -0.37274684201549374 0.46153169969114205 0.43040882407081466 -0.23395449789009468 0.35014673186737821 0.22531067965118223 0.2039204924730427 0.32189400477883445 0.17574148614037716;0.37476823932855202 0.11546669070368759 -0.42697958849517431 0.37463824728233669 0.18896936958088195 0.18544399738331774 -0.24707432492048281 -0.14104343087345794 0.12039635523989081 -0.26811616928246862 -0.13714994187791346 0.1999153343688525 0.066702891015596913 -0.1114256124701714 0.24349747505739053 -0.47482937127109714 -0.098103068685359104 0.2239529977248696 0.15363276001393569 -0.43615374081600777 0.27832482830345162 -0.13105214924599379 -0.087276269867450129 0.054836915925883564 -0.024874332288534421 -0.29015612700153071 0.40160084637794946 -0.21470551372720156 -0.12159309441070532 -0.22832695745345441 -0.4741872924427582 0.08010709366869595 -0.12709093111244582 0.36973518415031015;0.18370111112308748 -0.018898782741535738 -0.48518797982129946 0.38648694294404684 -0.35100960771497802 -0.13781366771574438 -0.20754819759285012 0.12883639256758114 0.5096664764071992 -0.024212741044357323 0.45052745554552653 0.20260353505680173 0.47228314972972085 -0.34187486274036016 0.65191151536123138 -0.50321162996530389 -0.47193644592037315 0.26423323011376681 -0.31612891144088717 0.37866202787421471 0.22892170749421537 0.36945359561830132 -0.11177925687789467 -0.14752554476440816 -0.50825852362040347 0.0013933572455434493 -0.59226035482801154 -0.68984679923098535 -0.4409676452727902 -0.23466028046389656 -0.00023069342390445073 0.14214091518537009 0.21732570202574586 -0.24084528223913657;0.16571676271596661 0.13496850985285166 0.5982752252872986 -0.53636450066464503 0.088728474732586174 0.067374497216185589 -0.18814899439421282 0.028559716948879071 0.19647378796700193 0.31069581039541155 -0.057785511053139595 -0.13580853602313606 -0.1574384402155081 0.12647300669219222 -0.50482015438014249 0.019589015246901886 0.29072712428189929 -0.038576615831027672 0.39003184476433883 -0.60246372479815968 0.19095848627343623 0.11064184137492346 -0.23926198257121573 0.28050298494740183 0.073503327329885154 -0.24521682276840695 0.47001328880601101 0.56538944455039464 -0.31120507963454896 -0.16173890361698254 -0.53568779513983777 -0.22691962213757588 0.017885878591252548 0.45929927005692966];
% Layer 2
b2 = [0.34791196117051243;0.032467720714978811];
LW2_1 = [0.79123446312813106 0.15350419751995401 0.11860830937971768 -0.59086739902228547 1.1296572286543383;-0.26622584110498199 -0.38821952149850331 -0.4888920545843492 1.247665944011028 -0.25135813450004862];
% ===== SIMULATION ========
% Format Input Arguments
isCellX = iscell(X);
if ~isCellX, X = {X}; end;
% Dimensions
TS = size(X,2); % timesteps
if ~isempty(X)
Q = size(X{1},2); % samples/series
else
Q = 0;
end
% Allocate Outputs
Y = cell(1,TS);
% Time loop
for ts=1:TS
% Input 1
Xp1 = mapminmax_apply(X{1,ts},x1_step1_gain,x1_step1_xoffset,x1_step1_ymin);
% Layer 1
a1 = tansig_apply(repmat(b1,1,Q) + IW1_1*Xp1);
% Layer 2
a2 = softmax_apply(repmat(b2,1,Q) + LW2_1*a1);
% Output 1
Y{1,ts} = a2;
end
% Final Delay States
Xf = cell(1,0);
Af = cell(2,0);
% Format Output Arguments
if ~isCellX, Y = cell2mat(Y); end
end
% ===== MODULE FUNCTIONS ========
% Map Minimum and Maximum Input Processing Function function y = mapminmax_apply(x,settings_gain,settings_xoffset,settings_ymin) y = bsxfun(@minus,x,settings_xoffset); y = bsxfun(@times,y,settings_gain); y = bsxfun(@plus,y,settings_ymin); end
% Competitive Soft Transfer Function function a = softmax_apply(n) nmax = max(n,[],1); n = bsxfun(@minus,n,nmax); numer = exp(n); denom = sum(numer,1); denom(denom == 0) = 1; a = bsxfun(@rdivide,numer,denom); end
% Sigmoid Symmetric Transfer Function function a = tansig_apply(n) a = 2 ./ (1 + exp(-2*n)) - 1; end
  1 个评论
Sean de Wolski
Sean de Wolski 2014-12-15
It's weird, you have to select the file, it will show up in the edit box and you then have to hit "attach". This happens to me all of the time.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by